Skillv0.1.0

ClawScan security

seisbench-model-api · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 14, 2026, 9:21 PM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill is an instruction-only overview of the SeisBench model API and its runtime instructions (pip install, using from_pretrained, reading waveforms) are coherent with that purpose.
Guidance: This skill is documentation/instructions for using the SeisBench Python library. Before installing or running examples: (1) Expect `pip install seisbench` and PyPI/network downloads; verify you trust PyPI and your environment's packages. (2) The `from_pretrained` calls will fetch pretrained model weights from the network and cache them locally—check available disk space and your network policy. (3) No secrets or credentials are requested by the skill. (4) If you run code produced by the agent, run it in a controlled environment (virtualenv/conda, sandbox) if you are concerned about package side effects. Overall the skill is internally consistent with its stated purpose.

Review Dimensions

Purpose & Capability: okThe name and description match the instructions: documentation and usage examples for SeisBench model API. Required capabilities (installing seisbench, loading/preprocessing waveforms, downloading pretrained weights) are appropriate for the stated purpose.
Instruction Scope: okSKILL.md contains examples that read seismic files (obspy.read), construct models, call annotate/classify, and documents that pretrained weights are downloaded and cached. It does not instruct reading unrelated system files, other credentials, or exfiltrating data to unexpected endpoints.
Install Mechanism: noteThere is no built-in install spec (instruction-only). The file recommends running `pip install seisbench`, and SeisBench's `from_pretrained` will download model weights at first use. This implies network downloads and disk writes but is expected for this skill; users should be aware of network and PyPI activity.
Credentials: okThe skill declares no required environment variables, no credentials, and no config paths. That is proportionate: using SeisBench needs no extraneous secrets. The documented behavior (downloading model weights) is consistent with no additional credentials being required.
Persistence & Privilege: okalways is false and there is no install-time or runtime behavior that attempts to persist or change other skills or system-wide agent configuration. The skill is user-invocable and can be invoked autonomously (platform default), which is expected for an instruction-only integration.